Generative AI Platform Case Study at RATP: When Run Becomes the Engine of the Product
summary
A RATP team presents a case study of building and operating a generative AI platform for multiple business uses, arguing that operations (Run) should drive the product lifecycle. They describe concrete practices: early risk reduction via documentation-focused opportunity studies and test strategies, a reusable platform of standards (patterns, prompts) to speed iterations, continuous enrichment of the platform from each use case, and ongoing evaluation/measurement to manage non-deterministic model quality.
AI-written summary · about
